Some more charming details about the recent Lancet study on the numbers of Iraqis killed in the last 18 months. This is excerpted from U.N. Chief's Remarks Irritate White House:

Of nearly 1,000 households visited by investigators, 808, representing nearly 8,000 people, took part. Each household was asked how many people lived in the home and how many deaths had occurred since January 2002, 15 months before the invasion. In most cases, death certificates were made available to the researchers.

Although the sample appears small in a country of roughly 25 million people, its size and the way it was carried out are considered standard for household surveys by social scientists working in developing countries. Moreover, because Fallujah, which was the site of major battles last April and has since been the target of numerous U.S. air strikes, was among the neighbourhoods surveyed, it was excluded from the final estimates because the death toll there was so high.

The investigation found that the most common causes of death before the invasion were heart attacks, strokes and chronic diseases. But after the invasion, violence had become the primary cause of death in Iraq, 58 times more likely than in the 15 months before the U.S.-led attack.

Of violent deaths, about 95 percent were attributed to bombing or fire from helicopter gunships. Most of the victims, according to the study, were women and children.

The estimated number killed is far beyond the 10,000 to 30,000 people suggested by independent groups, such as the Iraq Body Count project or the Brookings Institution, evoking incredulity by Brookings analyst Michael O'Hanlon, who called the findings ''preposterous and politically driven''.